A Pilot Study of Visual, Vocal, and Verbal Markers of PTSD

Current PTSD assessment methods often compromise precision for scalability, limiting clinical effectiveness. We propose a digital task-shifting framework that utilizes computational techniques to enhance assessment capacity while preserving clinician judgment. We implemented two structured trauma-activation tasks (trauma account and impact statement) via telehealth to elicit disorder-relevant behavioral signals, then quantified PTSD-related behaviors across verbal, vocal, and visual modalities. Parallel unimodal statistical analyses prioritized interpretability. Comparing 48 individuals with PTSD and 44 trauma-exposed controls helped separate the effects of trauma exposure from its psychopathological consequences. Bayesian generalized linear mixed models revealed significant main effects of group in the verbal and visual modalities and group-by-task interaction effects specifically in verbal and vocal modalities. These preliminary findings suggest that structured elicitation can reveal behavioral differences associated with PTSD, some consistent across tasks and some specific to one task.